Dual action of leptin on rest‐firing and stimulated catecholamine release via phosphoinositide 3‐kinase‐driven BK channel up‐regulation in mouse chromaffin cells

Abstract

Key points

• Leptin is an adipokine produced by the adipose tissue regulating body weight through its appetite‐suppressing effect and, as such, exerts a relevant action on the adipo‐adrenal axis.

• Leptin has a dual action on adrenal mouse chromaffin cells both at rest and during stimulation. At rest, the adipokine inhibits the spontaneous firing of most cells by enhancing the probability of BK channel opening through the phosphoinositide 3‐kinase signalling cascade. This inhibitory effect is absent in db^–/db^– mice deprived of Ob receptors.

• During sustained stimulation, leptin preserves cell excitability by generating well‐adapted action potential (AP) trains of lower frequency and broader width and increases catecholamine secretion by increasing the size of the ready‐releasable pool and the rate of vesicle release.

• In conclusion, leptin dampens AP firing at rest but preserves AP firing and enhances catecholamine release during sustained stimulation, highlighting the importance of the adipo‐adrenal axis in the leptin‐mediated increase of sympathetic tone and catecholamine release.

Abstract

Leptin is an adipokine produced by the adipose tissue regulating body weight through its appetite‐suppressing effect. Besides being expressed in the hypothalamus and hippocampus, leptin receptors (ObRs) are also present in chromaffin cells of the adrenal medulla. In the present study, we report the effect of leptin on mouse chromaffin cell (MCC) functionality, focusing on cell excitability and catecholamine secretion. Acute application of leptin (1 nm) on spontaneously firing MCCs caused a slowly developing membrane hyperpolarization followed by complete blockade of action potential (AP) firing. This inhibitory effect at rest was abolished by the BK channel blocker paxilline (1 μm), suggesting the involvement of BK potassium channels. Single‐channel recordings in ‘perforated microvesicles’ confirmed that leptin increased BK channel open probability without altering its unitary conductance. BK channel up‐regulation was associated with the phosphoinositide 3‐kinase (PI3K) signalling cascade because the PI3K specific inhibitor wortmannin (100 nm) fully prevented BK current increase. We also tested the effect of leptin on evoked AP firing and Ca²⁺‐driven exocytosis. Although leptin preserves well‐adapted AP trains of lower frequency, APs are broader and depolarization‐evoked exocytosis is increased as a result of the larger size of the ready‐releasable pool and higher frequency of vesicle release. The kinetics and quantal size of single secretory events remained unaltered. Leptin had no effect on firing and secretion in db^–/db^– mice lacking the ObR gene, confirming its specificity. In conclusion, leptin exhibits a dual action on MCC activity. It dampens AP firing at rest but preserves AP firing and increases catecholamine secretion during sustained stimulation, highlighting the importance of the adipo‐adrenal axis in the leptin‐mediated increase of sympathetic tone and catecholamine release.

Abbreviations

AHP

after‐hyperpolarization

AP

action potential

CC

chromaffin cells

DMEM

Dulbecco's modified Eagle's medium

HPA

hypothalamic‐pituitary‐adrenal

JAK

Janus kinase

MAPK

mitogen‐activated protein kinase

MCC

bovine chromaffin cell

mTORC1

mammalian target of rapamycin complex 1

nP_o

open probability of n channels

NPY

neuropeptide Y

ObR

leptin receptor

PI3K

phosphoinositide 3‐kinase

PKC

protein kinase C

POMC

pro‐opiomelanocortin

RRP

ready releasable pool

STAT3

signal transducer and activator of transcription 3

Introduction

Leptin is an adipokine produced by the fat tissue known for its inherent appetite‐suppressing effects. This 16 kDa protein is the product of the Ob (obese) gene and exerts its effects through a class I cytokine receptor (Tartaglia et al. 1995). Leptin receptors (ObRs) are associated with Janus kinases (JAKs), triggering a signalling cascade involving phosphoinositide 3‐kinase (PI3K) and, subsequent to the translocation of signal transducer and activator of transcription 3 (STAT3) into the nucleus, affecting gene transcription (Harvey et al. 2000; Zhao et al. 2000; Shanley et al. 2002 a). ObRs are widely expressed in numerous tissues: in the arcuate and ventromedial nucleus of the hypothalamus, for appetite inhibition (Halaas et al. 1995); in the hippocampus, where they regulate synaptic plasticity and neuroprotection (Huang et al. 1996; Gavello et al. 2012); and also in peripheral organs such as the chromaffin cells (CCs) of the adrenal glands, which constitute the main hub of the sympathetic nervous system (Hoggard et al. 1997; Glasow et al. 1998; Yanagihara et al. 2000).

The action of leptin on the hypothalamic‐pituitary‐adrenal (HPA) axis and the sympathetic/adrenomedullary (SS/AM) system comprises a central mechanism in the regulation of the body energy balance by affecting food intake, thermogenesis and energy expenditure (Halaas et al. 1995). When directly injected into the ventromedial hypothalamus, leptin increases sympathetic nerve activity and plasma levels of adrenaline and noradrenaline (Satoh et al. 1999). This action is mediated by PI3K and mammalian target of rapamycin complex 1 (mTORC1), whose modulation causes dramatic changes to sympathetic traffic, blood flow and arterial pressure (Harlan et al. 2013). Despite these important signalling details, little is known about the effects of leptin on peripheral targets such as the adrenal gland and sympathetic nerve endings, where leptin has been shown to increase catecholamine secretion (Satoh et al. 1999). Little is known also about the leptin‐mediated mechanisms downstream to PI3K that control sympathetic activity and circulating catecholamines (Harlan & Rahmouni, 2013).

Studies of the action of leptin on adrenal CCs suggest a potentiating effect of the adipokine on catecholamine secretion. In porcine CCs, leptin increases intracellular calcium levels by activating L‐ and N‐type Ca²⁺channels and inositol trisphosphate production (Takekoshi et al. 2001 a). In the same cells, leptin potentiates the synthesis of catecholamines by increasing tyrosine hydroxylase activity via protein kinase C (PKC) (Takekoshi et al. 2001 b) and mitogen‐activated protein kinase (MAPK) (Shibuya et al. 2002). The higher catecholamine synthesis is assumed to result in the increased release of adrenaline and noradrenaline (Takekoshi et al. 1999), although it is not known how leptin affects the dynamics of exocytosis in CCs. Only in PC12 cells is leptin shown to almost double the frequency of secretory events, as revealed by single‐cell amperometric recordings (Than et al. 2010). In this case, however, it remains to be determined whether this occurs through a leptin‐induced increase of Ca²⁺‐influx during exocytosis, an increased size of the ready releasable pool (RRP) of vesicles or a higher probability of vesicle fusion and catecholamine release. Concerning the effects on voltage‐gated Ca²⁺ channels, leptin either potentiates Ca_v1.3 L‐type calcium channels in pituitary gonadotrope LβT2 cells (Avelino‐Cruz et al. 2008) or inhibits the voltage‐gated Ca²⁺ channels of perifornical lateral hypothalamus neurons via the activation of JAK2 and MAPK (Jo et al. 2005).

Understanding the possible effects of leptin on adrenal CC activity and their output is further complicated by the observation that the adipokine activates the Ca²⁺‐ and V‐dependent BK channels of resting hippocampal neurons (Shanley et al. 2002 b; Gavello et al. 2012), thus supporting a neuroprotective role during brain damage (Zhang et al. 2007; Signore et al. 2008; Mancini et al. 2014). Given that BK channels are highly expressed in CCs (Prakriya & Lingle, 1999; Marcantoni et al. 2007; Marcantoni et al. 2010; Martinez‐Espinosa et al. 2014) and regulate the timing of CCs firing at rest and during prolonged stimulation (Vandael et al. 2010; Vandael et al. 2012; Vandael et al. 2015 a), it is evident that a leptin‐mediated increase of BK currents in CCs probably induces a depressive action on cell excitability. This would hardly fit with the expectation that leptin stimulates sympathetic nerve activity and catecholamine release (Halaas et al. 1995; Satoh et al. 1999; Harlan et al. 2013). Accordingly, we set up a detailed study in which we attempt to link the effects of leptin on mouse chromaffin cell (MCC) excitability to CC output in terms of catecholamine secretion.

In the present study, we show that acute application of leptin on MCCs induces a pronounced hyperpolarization driven by a PI3K‐mediated activation of BK channels via transmembrane ObRs. Cell hyperpolarization either reduces or stops the spontaneous firing and thus attenuates MCC excitability at rest. This result is in agreement with what has been observed in hippocampal neurons, where leptin blocks neuronal firing by activating Slo1 BK channels (Shanley et al. 2002 a; Gavello et al. 2012). Leptin nevertheless preserves robust cell firing adaptation during sustained cell depolarizations and potentiates MCC exocytosis, by increasing the frequency of vesicle release and the size of the RRP. The robustly adapted firing during prolonged stimulation and the increased rate of catecholamine release support the hypothesized potentiating action of leptin on sympathetic traffic, circulating catecholamines and blood pressure.

Methods

Ethical approval

All experiments were conducted in accordance with the European Communities Council Directive 2010/63/UE and approved by the Italian Ministry of Health (Authorization number 121/2015) and by the Local Bioethical Committee of the University of Turin. Every effort was made to minimize animal suffering and the number of animals used. For removal of tissues, animals were deeply anaesthetized by CO₂ inhalation and rapidly killed by cervical dislocation.

Isolation and culture of MCCs

MCCs were obtained from young male C57BL/6J and C57BL/6J db^–/db^– mice (3 months of age) (Taconic, Lille Skensved, Denmark), which were killed by cervical dislocation and cultured as described previously (Marcantoni et al. 2009; Vandael et al. 2012). After removal, the adrenal glands were placed in Ca²⁺‐ and Mg²⁺‐free Locke's buffer containing (in mm): 154 NaCl, 3.6 KCl, 5.6 NaHCO₃, 5.6 glucose and 10 Hepes (pH 7.3) at room temperature. The glands were decapsulated and the medullas were precisely separated from the cortical tissue. Medulla digestion was achieved after 30 min at 37°C in the enzyme solution [0.16 mm l‐cysteine, 1 mm CaCl₂, 0.5 mm EDTA and Dulbecco's modified Eagle's medium(DMEM)] containing 20 U ml⁻¹ of papain (Worthington Biochemical, Lakewood, NJ, USA) plus 0.1 mg ml⁻¹ of DNAse (Sigma, St Louis, MO, USA). The cell suspension was then washed two times with a washing solution (DMEM, 1 mm CaCl₂, 10 mg ml⁻¹ BSA) and resuspended in 2 ml of DMEM supplemented with 15% fetal bovine serum. Cells were plated in four‐well plastic dishes (glass Petri dishes for fluorescence measurements) treated with poly‐l‐ornithine (0.5 mg ml⁻¹) and laminin (10 μg ml⁻¹ in L‐15 carbonate) by placing a drop of concentrated cell suspension in the centre of one well. After 1 h, 1.8 ml of DMEM supplemented with 15% fetal bovine serum (Invitrogen, Grand Island, NY, USA), 50 IU ml⁻¹ penicillin and 50 μg ml⁻¹ streptomycin (Lonza, Allendale, NJ, USA), 10 μm cytosine b‐d‐arabino‐furanoside‐hydrochloride (Sigma) and 10 μm 5‐fluoro‐2′‐deoxyuridine (Sigma) was added to the wells. For intracellular Ca²⁺ measurements, cells were plated onto a glass Petri dish. Cells were then incubated at 37°C in a water‐saturated atmosphere with 5% CO₂ and used within 2–6 days after plating. All experiments were performed in accordance with the guidelines established by the National Council on Animal Care and were approved by the local Animal Care Committee of Turin University.

Voltage clamp and current clamp recordings

BK currents were measured in the perforated‐patch configuration using an Axon Multiclamp 700‐B amplifier and pClamp, version 10.0 (Molecular Devices, Sunnyvale, CA, USA) (Cesetti et al. 2003; Marcantoni et al. 2010). Patch pipettes were made of thin borosilicate glass capillaries (Kimax 51; Witz Scientific, Holland, OH, USA) and filled with an internal solution containing (in mm): 135 KAsp, 8 NaCl, 20 HEPES, 2 MgCl₂ and 5 EGTA (pH 7.4 with KOH) plus amphotericin B (Sigma). Amphotericin B was dissolved in dimethyl sulphoxide stored at −20°C in stock aliquots of 50 mg ml⁻¹ and used at a final concentration of 500 μg ml⁻¹. To facilitate the sealing, the pipette was first dipped in a beaker containing the internal solution and then backfilled with the same solution containing amphotericin B. EGTA was added in the pipette since increased the rate of success of pipette sealing and not for buffering intracellular Ca²⁺. Although no effort was made to test whether EGTA enters the cells, the selectivity of amphotericin B‐formed channels for small monovalent cations and non‐electrolyte molecules suggests negligible permeation of the Ca²⁺ chelator through the perforated‐patch (Kleinberg & Finkelstein, 1984; Horn & Marty, 1988; Rae et al. 1991). Pipettes with resistance of 1–2 MΩ were used to form giga‐seals. Series resistance of< 15 MΩ was compensated by 80% and monitored periodically during the experiment (Marcantoni et al. 2010). The external solution comprised standard Tyrode solution containing (in mm): 130 NaCl, 4 KCl, 2 CaCl₂, 2 MgCl₂, 10 HEPES and 10 glucose (pH 7.4 with NaOH). To avoid contamination by Na⁺‐activated K⁺ currents, TTX (300 nm) was added. Apamin (200 nm) was included to block SK currents when measuring BK currents and paxilline (1 μm) was added when SK currents were recorded (Vandael et al. 2012). Recombinant murine leptin (Peprotech, Rocky Hill, NJ, USA; catalogue number 450‐31; lot 30976; vial code C3009) was dissolved in BSA 1% and used at a final concentration of 1 nm. The adipokine was applied either acutely or incubated for 30 min at a concentration of 1 nm.

Current clamp experiments were performed in the perforated patch configuration, as described previously, using an Axon Multiclamp 700‐B amplifier and pClamp, version 10.0 (Marcantoni et al. 2009; Marcantoni et al. 2010; Vandael et al. 2012). The intracellular solution contained (in mm): 135 KAsp, 8 NaCl, 20 HEPES, 2 MgCl₂ and 5 EGTA (pH 7.4 with KOH). As an extracellular solution, we used normal physiological Tyrode solution containing (in mm): 130 NaCl, 4 KCl, 2 CaCl₂, 2 MgCl₂, 10 HEPES and 10 glucose (pH 7.4 with NaOH). Spontaneous activity was studied under control conditions and after the perfusion of the cells with leptin 1 nm.

Single‐channel recordings on perforated microvesicles

Single‐channel measurements were performed on perforated microvesicles. Perforated microvesicles were achieved by pulling off the patch pipette after amphotericin B lowered the access resistance below 30 MΩ. The advantage of using perforated microvesicles is the good voltage control achieved for the membrane patch at the same time as keeping the intracellular environment intact (Levitan & Kramer, 1990; Vandael et al. 2015 b). The technique is thus ideal for measuring single ion channel modulation mediated by intracellular second messengers (Levitan & Kramer, 1990; Maingret et al. 2008). For single‐channel experiments, glass micropipettes were pulled and forged to obtain pipette resistances in the range of 7–10 MΩ. Extracellular solution for single‐channel recordings was kept as described above, except that 4‐aminopyridine (5 mm) was included as a non‐specific blocker for K_V channels. Data were recorded at 20 kHz and filtered at 2 kHz with an eight‐pole low‐pass Bessel filter. Open and closed times were analysed using pClamp, version 10.0. The criteria for selecting the detection levels of channel opening and closing were similar to those described previously (Carbone & Lux, 1987).

Membrane capacitance changes

Ca²⁺ currents and the corresponding depolarization‐evoked capacitance changes were measured in the perforated‐patch configuration using an EPC‐10 double patch amplifier (HEKA Elektronic, Lambrecht, Germany) (Carabelli et al. 2003; Carabelli et al. 2007). MCCs were kept in saline solution, containing (in mm): 4 TEACl, 126 NaCl, 10 CaCl₂, 4 KCl, 2 MgCl₂, 10 glucose and 10 HEPES (pH 7.4 with NaOH). The internal solution contained in mM: 135 CsMeSO₃, 8 NaCl, 2 MgCl₂ and 20 HEPES (pH 7.3 with CsOH) plus amphotericin B (Sigma). Ca²⁺ currents were evoked by applying step depolarizations (100 ms in duration) from a holding potential of −70 mV to +10 mV. Superimposed on the square pulse, a sinusoidal wave function (1000 Hz, ± 25 mV) was applied to monitor membrane capacitance increases after the depolarizing step using the Lockin extension of Patchmaster software (HEKA Elektronic). Fast capacitive transients as a result of depolarizing pulses were minimized online by patch clamp analogue compensation. Series resistance was compensated for by 80% and monitored during the experiment. The amount of Ca²⁺ entering the cells during a depolarization (quantity of charge in pC) was calculated as the time integral of the Ca²⁺ current. Leptin was applied at a concentration of 1 nm for 30 min. All the experiments were performed at room temperature. Ca²⁺ currents were sampled at 10 KHz and low‐pass filtered at 2 KHz.

Analysis of the RRP

A dual‐pulse protocol consisting of two consecutive pulses of depolarization was used to calculate the size of the RRP (Gillis et al. 1996; Gavello et al. 2013). The two depolarizations of 100 ms in duration were set at 0 and +5 mV at short interval (100 ms), with the aim of applying two equivalent current injections. ΔC ₁ and ΔC ₂ indicate the capacitance changes measured after the pulses. Mean values of upper limit of the RRP (B _max) were calculated from (ΔC ₁ + ΔC ₂)/(1 − (ΔC ₂/ΔC ₁)²), whereas the probability of release (p) was obtained from 1 − (ΔC ₂/ΔC ₁). Before testing, leptin was dissolved in Tyrode solution and applied for 30 min at a concentration of 1 nm.

Amperometric detection of exocytosis

Amperometric recordings were performed by means of commercial carbon fibres (ALA Scientific Instrument Inc., Westbury, NY, USA) as described previously (Carabelli et al. 2007; Carabelli et al. 2010). Before each trial, carbon fibres were cut at an angle of 45 deg, inserted into the head‐stage of a EPC‐10 HEKA amplifier, polarized to +800 mV and, finally, positioned next to the cell membrane. MCCs were kept in saline solution, containing (in mm): 130 NaCl, 2 MgCl₂, 10 glucose, 10 HEPES, 10 CaCl₂ and 4 KCl (pH 7.4 with NaOH). Then exocytosis was stimulated by using a KCl‐enriched solution, containing (mm): 100 NaCl, 2 MgCl₂, 10 glucose, 10 HEPES, 10 CaCl₂ and 30 KCl. Amperometric currents were sampled at 4 kHz, low‐pass filtered at 1 kHz and monitored over 120 s. This same stimulation protocol has been followed both for control cells and after incubation with leptin 1 nm (30 min). Recordings were analysed using IGOR macros (Wave‐Metrics, Lake Oswego, OR, USA) (Carabelli et al. 2007). For each amperometric spike, we quantified: (1) the maximum oxidation current (I _max) as the height of each spike; (2) the vesicular charge content (Q) as the time integral of the current, as well as its cubic root (Q ^1/3), which is proportional to the vesicle diameter; (3) the rise‐time of the spike 25–75% (m); (4) the width of the spike at half of its height (t _1/2); and (5) the time to the peak (t _p). Before testing, leptin was dissolved in Tyrode solution and applied for 30 min at a concentration of 1 nm.

Statistical analysis

Data are reported as the mean ± SEM for the number of cells (n). Statistical significance was calculated either using Student's t test or ANOVA followed by the Bonferroni post hoc analysis. P < 0.05 was considered statistically significant.

Results

Leptin decreases spontaneous firing of MCCs through BK channel activation

To explore the effect of leptin on MCC excitability, we compared the spontaneous activity of MCCs under control conditions and after acute application of leptin (1 nm) by means of current clamp perforated‐patch recordings (Marcantoni et al. 2009; Marcantoni et al. 2010). Under control conditions, 80% of MCCs exhibited spontaneous activity for the entire duration of the recording (10–15 min) with a mean firing frequency of 0.83 ± 0.10 Hz and a mean membrane resting potential (V _rest) of –43.3 ± 0.6 mV. Acute application of leptin (1 nm) hyperpolarized the cells up to –59.1 ± 1.7 mV (P < 0.001) and was followed by inhibition of spontaneous activity (Fig. 1 A). Before cells stopped firing, leptin boosted the after‐hyperpolarization (AHP) peak from –50.5 ± 0.5 mV to –52.5 ± 0.8 mV (n = 12, P < 0.05) (Fig. 1 A, inset), as estimated from the 0 mV line. The AP peak amplitude had a tendency to increase by up to 20% (P > 0.05), whereas the cell progressively hyperpolarized by the effect of leptin. This was most probably the result of an increased availability of Na_v channels recovering from inactivation during the leptin‐induced hyperpolarization (Vandael et al. 2015 a; Vandael et al. 2015 b)

Figure 1. Current clamp recordings .

A,spontaneous APs recorded in current clamp mode from a representative MCC before, during and after leptin application (1 nm). Top right: overlapped APs shown at an expanded time scale to allow comparison of APs for the control and with leptin. B,spontaneous activity followed by application of paxilline (1 μm) and paxilline (1 μm) + leptin (1 nm). C,spontaneous activity followed by UCL 1684 (200 nm) and leptin + UCL 1684 application. D,spontaneous activity followed by leptin (1 nm) and, subsequently, leptin + wortmannin (100 nm) application. E,effect of leptin on db^–/db^– MCCs lacking ObRs. F and G,mean baseline and frequency values in wild‐type and db^–/db^– mice, under control conditions (black bar) and after leptin (red), paxilline (light grey) or wortmannin (dark grey).

Because of the role of BK currents in shaping the AHP phase of the AP (Marcantoni et al. 2010; Vandael et al. 2012), as well as the potentiating effects of leptin on BK channels in hippocampal neurons (Shanley et al. 2002 b), we tested whether BK channels were involved in the effect ascribed to leptin on MCC firing. We found that paxilline (1 μm), a specific BK channel blocker that binds almost exclusively to the closed‐channel conformation (Zhou & Lingle, 2014), abolished the inhibitory effect of leptin on AP firing (Fig. 1 B and F), this suggests an increased BK channel tone in the leptin‐mediated suppression of MCC firing at rest. We previously reported that the effect of paxilline alone is a 2.4 mV depolarization in MCCs that, in some cells, is not even evident (Marcantoni et al. 2010). To avoid misinterpretations, we thus conducted experiments where we inverted the order of leptin vs. paxilline application. In all cases, BK channel block by paxilline abolished the leptin‐induced effects on membrane resting potential and firing frequency (Fig. 1 B, D, F and G). Furthermore, we tested for a possible involvement of SK channels on the action of leptin. Application of the SK selective inhibitor UCL 1684 (200 nm) did not prevent leptin from hyperpolarizing the resting membrane potential. UCL caused a slight depolarization (from –43 to –39 mV; P > 0.05) with an expected increase of firing frequency (Vandael et al. 2010) followed by a comparable hyperpolarization when leptin was added and V _rest returned to –43.6 ± 1.8 mV (n = 4) (Fig. 1 C). Because leptin induces PI3K activation in hippocampal and hypothalamic neurons (Shanley et al. 2002 b; Hill et al. 2008), we tested whether this occurred on CCs as well. We found that, although leptin alone hyperpolarized the resting membrane potential, co‐application of leptin and wortmannin (100 nm), a specific PI3K blocker, prevented this effect, restoring the control baseline values (–47 ± 3 mV) (Fig. 1 D). Finally, we set up experiments aiming to confirm that leptin exerted its action in a specific way by binding to ObR. As a result of a lack of specific ObR blockers, recordings were performed on db^–/db^– MCCs lacking ObRs. As shown in Fig. 1 E, leptin had no effects on the resting potential of knockout MCCs (V _rest = –45.6 ± 0.7 mV for the control vs. –45.1 ± 0.6 mV with leptin, n = 8; P ≥ 0.05). Although a tendency towards an increased spontaneous firing rate, associated with a decreased AP amplitude, was evident, this parameter was not found to be significantly affected (0.8 ± 0.2 Hz for the controlvs. 1.3 ± 0.3 with leptin, n = 8) (Fig. 1 F and G). This confirms that the inhibitory action of leptin on spontaneously firing MCCs is predominantly mediated by ObRs and that, in the absence of these receptors, leptin exerts minor opposite effects on MCC excitability. These effects were not statistically significant and thus not investigated further in the present study.

Leptin increases BK current and shifts BK channel activation to more negative potentials

As reported above, leptin alters MCC excitability through BK channel activation. To obtain conclusive evidence in this respect, we performed a set of voltage clamp experiments in the perforated‐patch configuration. BK currents were elicited by a 50 ms pre‐pulse to 0 mV aiming to inject Ca²⁺ions and were measured during the subsequent test pulse, ranging from potentials of between –60 and +120 mV in steps of 20 mV. Currents were measured under control conditions and again after 3 min of 1 nm leptin application. Finally paxilline was applied to block the BK component. Subtraction of the traces measured in paxilline from control/leptin traces ultimately led to the BK current (Fig. 2 A). Accordingly, the voltage‐dependence of BK conductance, g_BK(V), was measured and data were fit by a Boltzmann function. We could clearly observe that, as expected, leptin led to a significant (38%) increase in maximal conductance compared to control (n = 16; P < 0.05, paired Student's t test). Moreover, we noted a 12 mV leftward shift in the half‐maximal activation that is similar to the action of some BK channel opener on g_BK(V) (Ha et al. 2006; Bentzen et al. 2014). We also found that leptin does not alter significantly the fraction of inactivating vs. non‐inactivating BK currents (data not shown). Although the short duration of the pre‐pulse used (50 ms) does not allow a rigorous quantitative analysis, this may suggest no specific action on either one of the two types of BK channels (fast and non‐inactivating) normally expressed in MCCs (Marcantoni et al. 2010; Martinez‐Espinosa et al. 2014).

Figure 2. Effects of leptin on BK and SK currents in MCCs .

A,right: BK currents measured in accordance with the protocol shown. Current peak amplitudes were used to calculate the conductance (g_BK) (left). (g_BK) was calculated as I _peak/(V − E _rev). Data were fit by a Boltzmann equation with mean V _1/2 = –34.2 mV (control) and V _1/2 = –21.9 mV (leptin). B and C, dependence of BK currents on pre‐conditioning voltage and duration, respectively. Different colours in (B) indicate test depolarizations to different voltages. D,slow tail currents elicited by the protocol shown in the right‐hand inset. Slow tails reflect SK current deactivation. Different colours were used to identify traces during Ca²⁺‐loading steps of different duration. Data in (C) and (D) were fit with a single exponential function: $I\left(t\right)=A{e}^{-\mathrm{t}/\tau }+C$

Next we tested for the Ca²⁺‐dependence of BK channel activation by varying the pre‐pulse voltage of 50 ms in duration between –60 and +140 mV (Fig. 2 B). BK current peaks measured during the test pulse at 120 mV clearly followed a bell‐shaped distribution, as expected for Ca²⁺‐dependent BK channels (Marty & Neher, 1985; Neely & Lingle, 1992). Leptin was found to increase the peak BK current by ∼40% (P < 0.001). No apparent shift of the peak value was observed, indicating that this effect cannot be attributed to a specific effect on L‐ or non‐L‐type calcium channels. By varying the pre‐step duration (from 5 ms up to 100 ms, in steps of 5 ms), we found, for each step duration, higher BK peak values with leptin compared to control. Data were fit by first‐order exponential functions and, although not significant, time constants revealed a trend to be faster after leptin application compared to control (Fig. 2 C). Although the current clamp data clearly indicate that SK currents are not affected by leptin (Fig. 1 C), we tested for the effect of leptin on macroscopic SK currents (Fig. 2 D). The protocol used consisted of Ca²⁺‐loading steps of increasing duration (from 5 ms up to 285 ms) followed by a 3 s return step to –120 mV. This typically gives rise to slow tail currents that are effectively blocked by 200 nm apamin or UCL 1684 (Vandael et al. 2012). No significant differences were found for either of the Ca²⁺ loading step durations when comparing control slow tail current amplitudes with those measured after leptin application (data not shown).

Leptin acts through the ObR receptor/PI3K pathway and does not affect K_v currents

We next performed a series of experiments to test for the specificity of the leptin effect and for the role of Ca²⁺ released from intracellular stores in the previously described phenomenon. Accordingly, a protocol consisting of a depolarizing step from –70 mV to +90 mV was applied for 200–500 ms with an intersweep interval of 5 s (Fig. 3 A). Application of leptin (1 nm) induced an increase of ∼30% of steady‐state K⁺ outward current (3.5 ± 0.4 nA control and 4.9 ± 0.5 nA with leptin; n = 9; P < 0.001, ANOVA Bonferroni post hoc), which was blocked by 1 μm paxilline (3.3 ± 0.3 nA; P > 0.05, ANOVA Bonferroni post hoc) (Fig. 3 A), suggesting a possible increase of BK currents. To test whether leptin could affect K_V currents, 200 μm Cd²⁺ was added in combination with paxilline to exclude any contribution of Ca²⁺‐activated K⁺ channels. In the presence of Cd²⁺ and paxilline, leptin did not lead to any increment of current, indicating clearly that K_V currents are left unaltered (Fig. 3 B). The same observation was made in six MCCs. Paxilline was added in combination with Cd²⁺ because BK channels are voltage‐dependent and can be activated by consistent depolarizations alone (Magleby, 2003). This experiment unequivocally confirms that Ca²⁺ from an external source is indispensible for BK channel activation.

Figure 3. Voltage clamp recordings of BK currents .

A,representative potassium currents (K_v + BK) recorded under control conditions (Ctrl), as well as during leptin (1 nm, lep) and paxilline application (1 μm, pax). The protocol is shown at the bottom. Right: mean currents normalized to the control value. The intersweep interval was fixed at 5 s. Inset: mean value of normalized current for the control (black), leptin (red) and paxilline (blue). B–D,potassium current traces in the presence of Cd²⁺ (200 μm, black), EGTA‐AM (20 μm, black) and wortmannin (100 nm, grey), a specific blocker of PI3K. E,representative recordings in db^–/db^– mice. Data were tested for statistical significance by one‐way ANOVA followed by a Bonferroni post hoc test.

ObR signalling typically leads to Ca²⁺‐induced Ca²⁺ release from intracellular stores (Takekoshi et al. 2001 a). Widely used external Ca²⁺ chelators such as EGTA and BAPTA limit Ca²⁺ diffusion to micro‐ and nanodomains, respectively (Fakler & Adelman, 2008). Previous work has provided evidence in favour of a strong coupling between L‐type calcium channels (Ca_V1.3) and BK channels in MCCs (Marcantoni et al. 2010). Although improbable (given the absence of BK currents in the presence of Cd²⁺), we tested whether Ca²⁺ deriving from intracellular stores by diffusion could lead to the leptin‐driven increase in BK channel tone. We thus performed the experiment in the presence of EGTA‐AM (20 μm), which is able to effectively buffer the intracellular Ca²⁺ levels of MCCs, regardless of the presence of EGTA in the patch pipette (Vandael et al. 2012). The recordings shown in Fig. 3 C demonstrate that EGTA‐AM loading of cells (40 min) does not interfere with the leptin‐mediated increase of the BK current (2.5 ± 0.2 nA for EGTA; 3.1 ± 0.3 nA for leptin; n = 7; P < 0.05). This suggests that Ca²⁺ released from intracellular stores does not have a crucial role in leptin‐mediated activation of BK channels (Fig. 3 C). The alternative possibility that the slow Ca²⁺buffer EGTA fails to prevent the effects of leptin because of the close proximity of Ca²⁺ stores to BK channels is improbable. BK channels in MCCs are predominantly activated by voltage‐gated Ca²⁺ channels (Fig. 2 B), which are tightly coupled to them (Marcantoni et al. 2010; Vandael et al. 2010; Martinez‐Espinosa et al. 2014).

As noted earlier above, PI3K has a pivotal role in the leptin associated signalling pathway in different cellular models. Wortmannin (100 nm) indeed reversibly blocked the leptin‐induced increase of BK current (4.2 ± 0.7 nA control, 5.7 ± 0.8 nA leptin, 3.3 ± 0.6 nA wortmannin and leptin; n = 6; P < 0.001, ANOVA Bonferroni post hoc) (Fig. 3 D). This clearly confirms the involvement of PI3K in the leptin‐signalling cascade in MCCs. These data are in line with the previously described effects of leptin on MCC spontaneous firing obtained in current clamp mode. Again, the effects were mediated by ObRs because leptin failed to affect BK currents in db^–/db^– mice (7.0 ± 1.2 nA control vs. 6.7 ± 1.1 nA with leptin; n = 9; P > 0.05) (Fig. 3 E). The contribution of BK channels to total K⁺‐driven outward currents changed remarkably in experiments on wild‐type MCCs from a minimum of 9% (Fig 3 A) to a maximum of 48% (Fig 3 D). This may derive from the non‐optimal conditions of Ca_V channel activation at potentials as high as +90 mV. When wild‐type data were pooled (n = 19) and compared with those obtained from db^–/db^– MCCs, we found that paxilline blocked 29 ± 4.3% and 54 ± 2.8% of the total outward current, respectively (P < 0.01). This could suggest that db^–/db^– MCCs might express higher densities of BK channels compared to wild‐type, presumably as a compensatory mechanism. This hypothesis, however, was not investigated further.

Single BK channel open probability is potentiated by leptin

To confirm that BK channels are directly affected by leptin without an indirect effect on Ca²⁺ influx, we performed single‐channel experiments on isolated ‘perforated microvesicles’(Vandael et al. 2015 b). The reasons for using perforated microvesicles are: (1) the excellent voltage control; (2) the possibility of directly applying specific blockers to identity the channel; and (3) the preservation of the internal cell environment, which allows testing for single ion channel modulation under alsmost physiological conditions. In a first set of experiments, we clamped the voltage at 0 mV in gap‐free mode (continuous recording). Under this condition, Ca²⁺ channels are typically inactivated, implying that we can test for direct BK channel modulation by leptin, independently of intracellular Ca²⁺ increase. Because of the experimental conditions, it is extremely hard to obtain perforated microvesicles containing single BK channels; thus, accurate analysis of the single‐channel dwell time and open probability is rather limited in perforated microvesicles. However, we found that BK channel openings became drastically more frequent after leptin application. As shown in Fig. 4 A, leptin (1 nm) increased the open probability (nP _o) of the BK channels present in the isolated membrane patch from 0.43 ± 0.15 to 2.2 ± 0.4 (n = 8; P < 0.01). This potentiating effect was removed by paxilline (1 μm) restoring the nP _o values to 0.07 ± 0.04 in presence of leptin. This indicates unequivocally that the recorded unitary outward currents were carried by BK channels. Current amplitudes could be fit with a single Gaussian function under control conditions centred at 4.5 pA (Fig. 4 D). After leptin application, two clear separate peaks emerged and were fit by two Gaussian functions centred at 4.9 and 9.3 pA with minimal overlap (Fig. 4 D), suggesting simultaneous multiple openings. A separate set of experiments was subsequently set‐up to check for any eventual leptin‐induced changes in BK single‐channel conductance. Membrane potential was raised from –30 mV up to +70 mV in steps of 10 mV (Fig. 5 A). The slope of the linear regression between the amplitude of the current vs. the voltage was the same for the control (filled squares) and with leptin (open circles) (101.4 pS for the control vs. 100.8 pS with leptin), confirming that leptin increases BK single‐channel open probability but not the single‐channel conductance (Fig. 5 B and C). In the presence of either paxilline or wortmannin, leptin did not succeed in increasing BK channel open probability because the nP _o values are comparable to controls (Fig. 5 C). Thus, leptin is acting on BK channels through a PI3K‐mediated pathway.

Figure 4. Single BK channel recording .

A,representative single‐channel recording obtained by means of the ‘perforated microvesicle’ technique in the presence of leptin (1 nm, lep) and paxilline (1 μm, pax). B,open probability of BK channels under control conditions, with leptin and paxilline. C,enlargement of the trace in (A) showing BK single‐channel openings under control conditions and in the presence of leptin or paxilline. D,distributions of BK channel amplitudes under control conditions (left) and in the presence of leptin 1 nm (right).

Figure 5. Unitary BK channel conductance .

A,representative single BK channel recordings at different potentials under control conditions (left) and during acute application of leptin 1 nm (right panel). B,unitary conductance of controls vs. leptin. There is no significant difference between the slope of the linear regression lines (101.4 pS for the control vs. 100.8 pS for leptin). C,open probability of BK channels under control conditions, with leptin 1 nm, leptin + paxilline (1 mm) and wortmannin + leptin. Application of leptin induces a significant increase of BK channel open probability (**P < 0.01 with one‐way ANOVA followed by Bonferroni post hoc test).

Leptin broadens the AP width and preserves the ability to generate AP trains of lower frequency during sustained depolarization

Having shown that leptin inhibits the spontaneous firing of MCCs by up‐regulating BK channels through a PI3K‐mediated pathway, we tested whether leptin had also a similar inhibitory action on AP firing during sustained depolarizations. We tested this by exposing the cells to chronic incubation of the adipokine (30 min, 1 nm) for three reasons: (1) to allow leptin to reach its maximal regulatory effects that usually require 5–10 min (Fig. 3); (2) to assay the effects of leptin on AP shape and spike frequency adaptation during prolonged depolarizations mimicking MCCs responses to sustained sympathetic stimulation; and (3) to compare the effects of leptin on cell excitability with the changes of catecholamine secretion that are estimated in separate groups of control and leptin‐treated MCCs (see below). We first checked whether leptin incubation altered V _rest and the input cell resistance (R _input) of MCCs at rest. Leptin had a hyperpolarizing effect on V _rest similar to that observed during acute application (Fig. 1 A and F). V _rest was –47.9 ± 2.9 mV (n = 8) at control and –59.1 ± 2.4 mV (n = 8; P < 0.05) after leptin incubation. The R _input, estimated using hyperpolarizing current pulses of –2 to –6 pA, had a tendency to decrease in cells treated with leptin (3.8 ± 0.3 GΩ, n = 10 with leptin vs. 4.1 ± 0.6 GΩ, n = 7 at control), although the reduction was not statistically significant (P > 0.05, unpaired Student's t test). This suggests that the expected decrease of R _input caused by the increased BK channel conductance is a small detectable percentage (∼10%) possibly resulting from a compensatory decrease of L‐type and BK channel conductance at more hyperpolarized potentials (Marcantoni et al. 2010; Berkefeld & Fakler, 2013).

Leptin‐treated cells responded well to sustained depolarization. During 6 pA current injection, most control and leptin‐treated MCCs fired one or more APs (Vandael et al. 2012). The percentage of firing was undistinguishable at this threshold stimulation: 9 out of 11 control cells and 13 out of 16 leptin‐treated cells fired multiple APs (P > 0.65, calculated with a chi‐squared test). When current intensity was raised, firings became sustained and marked differences were noted between control and leptin‐treated MCCs. Control cells had higher initial (f _o) and steady‐state firing frequency (f _ss): mean f _o was 15.6 Hz at control during 18 pA current injection (n = 11) and decreased to 12.2 Hz with leptin (n = 16; P < 0.05), mean f _ss was 3.7 Hz at control and decreased to 2.6 Hz with leptin (P < 0.05) (Fig. 6 A), suggesting stronger adaptation of leptin‐treated cells. The initial APs (Fig. 6 B) had indistinguishable amplitude, overshoot and AHP values but were slightly broader in the presence of leptin compared to control APs: mean half‐width was 3.2 ms (n = 11) at control and 4.0 ms with leptin (n = 16) (P < 0.05) (Fig. 6 B). Broadening was even more evident on the APs at the end of the pulse. The last AP of control and leptin‐treated cells had mean half‐widths of 5.9 ms (n = 11) and 8.9 ms (n = 16) (P < 0.01), respectively (Fig. 6 B, circles). These characteristics were more apparent when comparing the phase plane plots (Fig. 6 C). The ‘adapted’ APs after the first two spikes had remarkably different δV/δt vs.V trajectories. The ‘adapted’ APs for the control reached almost equal maximal positive and negative rates of rise (mean +15.7 V/s and –12.6 V/s; n = 5) at almost the same voltage (–8 mV; black arrows). Those of leptin‐treated cells instead had their maximal positive slope comparably high with respect to controls (mean +13.9 V/s; downward red arrow) but had lower negative slopes soon after the peak (mean –4.9 V/s, n = 6, P < 0.01) (upward red arrow). This indicates a net broadening of APs after chronic leptin treatment. The apparent discrepancy that broadening of spontaneous APs at rest was not observed (Fig. 1 A) is probably a result of the shorter exposure time of cells to leptin during acute applications (a few minutesvs. 30 min).

Figure 6. Leptin preserves MCC excitability during sustained depolarization and does not alter Ca²⁺ current densities .

A,spike frequency adaptation is preserved in leptin‐treated MCCs. Representative current clamp recordings for the control (left) and after leptin incubation (1 nm for 30 s; right) are shown after 6, 12 and 18 pA current injection from V _h = −70 mV. fo, instantaneous firing frequency at onset (first interspike interval) and steady‐state‐frequency; fss, instantaneous firing frequency at steady‐state (last interspike interval) (*P < 0.05; right‐hand inset). B,leptin causes a broadening of APs during evoked stimulations. Three APs taken from (A) control and with leptin, labelled with stars, triangles and circles, are overlapped. Mean half‐width values of the first and last APs are shown in the right‐hand inset (*P < 0.05; **P < 0.01). C,phase plane plots (δV/δt vs.V) from APs measured for the control (black) and after leptin treatment (red), respectively. V _thresh at control is defined as the point where δV/δt reaches 4% of its maximal value (indicated as δV/δt) and was not significantly different with leptin. Inset to the right: mean ± SEM values of the positive and negative δV/δt _max for the control and with leptin calculated at the points indicated by the arrows (**P < 0.01). D,representative Ca²⁺ currents for the control (black trace) and after leptin treatment (red trace) evoked by a ramp command, from –50 to +60 mV (0.87 V/s) in 10 mm extracellular Ca²⁺. The approximate potential of half‐activation is +8 mV and +23 mV more depolarized with respect to 5 and 2 mm Ca²⁺(Marcantoni et al. 2009; Marcantoni et al. 2010). Inset: mean values of Ca²⁺ current densities for the control and after leptin treatment. Two representative Ca²⁺ currents are shown to the right for the control (black trace) and after leptin treatment (red trace) measured during a depolarizing step of 50 ms to +10 mV in 2 mm Ca²⁺. Mean half‐time‐to‐peak (t _p1/2), percentage of inactivation (% inact) and slope factor (k) were not statistically different: t _p1/2 = 1.13 ms, % inact 0.22, k = 7.7 mV for the control (n = 17); t _p1/2 = 1.05 ms, % inact 0.28 and k = 7.9 mV for leptin‐treated cells (n = 16). E,representative Na_v and Ca_v currents under control conditions (black trace) or after leptin treatment (red trace) recorded at +10 mV in 10 mm Ca²⁺. To better resolve the fast inactivating Na_v current, only the early part of the recording is shown. Inset: mean values of the normalized Na_v current peak for the control (black) and after leptin treatment (red).

To test the origin of the broadening, we first checked whether leptin treatment affected the size, kinetics and voltage‐dependence of Ca²⁺ currents. Measurements were performed in 10 mm external Ca²⁺ to enable direct comparison with the effects on exocytosis that are usually studied at high Ca²⁺ (Carabelli et al. 2007; Marcantoni et al. 2009). Figure 6 D shows that leptin did not alter the size of Ca²⁺ current densities activated with ramp commands of 0.87 V/s (mean peak 30 pA/pF in control, n = 19 and 27 pA/pF with leptin, n = 10; P > 0.05) (Fig. 6 D, inset). Leptin caused mainly a slight voltage shift of –6 mV of the I/V curve (V _1/2 was –9.9 ± 2.1 mV, n = 6 for the control vs. –15.8 ± 1.3 mV, n = 9 with leptin; P < 0.01) with no changes of voltage‐dependence and activation/inactivation kinetics tested with square pulse depolarizations of 50 ms at +10 mV (Fig. 6 D, right). Finally, we tested whether leptin affected the size of Na⁺ currents by applying a depolarizing pulse of 100 ms at +10 mV when Na_V and Ca_V channels were both available. As shown in Fig. 6 E, leptin had no effect on the early peak amplitude of the current associated with Na_V channels (mean 110 pA/pF in controls vs. 119 pA/pF in leptin‐treated cells; P > 0.05) (Fig. 6 E).

Taken together, these data suggest that leptin preserves the ability of MCCs to fire and quickly adapt to a lower rate of firing mainly because of the enhanced activity of BK channels. This prolongs the repolarization phase between consecutive APs and, together with the increased Ca_V channels availability at potentials< 0 mV, boosts the effective recruitment of Na_V and Ca_V channels to broaden and stabilize the size of the ‘adapted’ APs. The final result is an increased Ca²⁺loading that enhances catecholamine release during sustained stimulation.

Leptin increases the secretory response of CCs without affecting Ca²⁺‐entry

Given the unexpected firing phenotype and the larger AP widths induced by leptin, we were eager to investigate the effect of leptin on evoked catecholamine release. To evaluate the effect of leptin on the secretory apparatus of CCs, we used two complementary approaches. First, we measured membrane‐capacitance changes (under voltage clamp conditions) to monitor the coupling between Ca²⁺‐influx and the quantity of depolarization‐evoked exocytosis. These recordings provide crucial information about the size of the readily releasable pool of secretory granules. Our capacitance measurements were than complemented by amperometric recordings to evaluate the kinetics and the size of single exocytotic events (Carabelli et al. 2007).

In a first set of experiments, the membrane capacitance increases (ΔC) were measured after 100 ms depolarization to +10 mV, with the aim of investigating the Ca²⁺‐dependent secretion of the whole cell (Fig. 6 A). Experiments were performed under stationary flux conditions (Carabelli et al. 2003; Giancippoli et al. 2006; Carabelli et al. 2007), by comparing control cells with those exposed to leptin (1 nm) for 30 min.

The main effect of leptin exposure was an increase of the overall exocytosis without altering the amount of Ca²⁺‐entry. As shown in Fig. 7 A and B, depolarization‐evoked exocytosis leads to a mean ΔC increase of 51 ± 5 fF (n = 34) under control conditions (Gavello et al. 2013), whereas, during 30 min of exposure to leptin, exocytosis increased to 84 ± 12 fF (n = 35 cells) (P < 0.05, one‐way ANOVA followed by Bonferroni post hoc analysis). The increased secretory response was not induced by an altered Ca²⁺ influx. Indeed, the quantity of charge (Ca²⁺ ions) entering during the depolarizing pulse through open voltage‐gated Ca²⁺ channels was not affected by leptin. For control cells, the mean quantity of charge was 3.1 ± 0.4 pC/pF and, after leptin treatment, remained almost unaltered (3.8 ± 0.6 pC/pF; P > 0.1, one‐way ANOVA followed by Bonferroni post hoc analysis) (Fig. 7 B). To ensure that the potentiating effects on catecholamine secretion were mediated by the ObR, we tested whether leptin had no action in db^–/db^– mice. As shown in Fig. 7 A, leptin had no sizeable effect on catecholamine secretion in db^–/db^– MCCs. ΔC was 29 ± 5 fF for the control and 31 ± 5 fF with leptin (n = 22; P ≥ 0.1, one‐way ANOVA followed by Bonferroni post hoc analysis), thus confirming that the observed effects on secretion are mediated by the ObR chain activated pathway.

Figure 7. Membrane capacitance measurements .

A,representative secretory measurements in wild‐type (top) and knockout (KO) (bottom) preparations. Lower traces represent Ca²⁺ currents and upper traces represent the depolarization‐evoked secretory responses (measured at +10 mV) for the control and for leptin‐treated (1 nm, 30 min) CC. B,quantity of charge and mean capacitance changes (ΔC) of control vs. treated cells in wild‐type and KO mice. ΔC is evaluated at 500 ms after the pulse (*P < 0.05; using ANOVA followed by Bonferroni post hoc analysis). C, representative RRP measurements under control conditions and after application of leptin 1 nm (30 min). The RRP size was evaluated by means of the double‐pulse protocol. Right: mean estimated values of the maximal RRP.

To further investigate the effects of leptin on MCCs secretory pathways, we estimated the size of the RRP using the double‐pulse protocol described by Gillis et al. 1996 (see also Carabelli et al. 2003; Gavello et al. 2013).The dual‐pulse protocol was designed to elicit secretion with two consecutive pulses, taking care that the two Ca²⁺ injections applied in rapid succession were comparable. For this reason, the two depolarizing steps were set at 0 and +5 mV (100 ms). From the sum and the ratio of the two consecutive capacitance increases, ΔC ₁ and ΔC ₂ (Gillis et al. 1996), we estimated the maximum size of the RRP (B _max) from the equation B _max = (ΔC ₁  + ΔC ₂)/[1 − (ΔC ₂/ΔC ₁)²]. As shown in Fig. 7 C, the maximal RRP size increased from 123 ± 13 fF (n = 18) for the control to 185 ± 18 fF (n = 23, P < 0.05) after leptin exposure (1 nm for 30 min).

Leptin increases the frequency of quantal events without affecting their size and time course

To assess whether leptin could also alter the kinetics and amplitude of single quantal events, we measured the amperometric spikes after exposure to leptin (1 nm for 30 min). Amperometric currents associated with the oxidation of released catecholamines from chromaffin granules were evoked by a KCl‐enriched external solution (see Methods) and measured for 120 s. Figure 8 A shows two examples of amperometric recordings for the control (Fig. 8 A, top) and leptin‐treated MCCs (Fig. 8 A, bottom). In both cases, amperometric signals occurred in bursts, although the single event had similar time course and amplitude (Fig. 8 A, inset on the right). The only notable change was in the rate of occurrence of the events whose frequency increased from 0.51 ± 0.05 Hz for the control to 0.8 ± 0.1 Hz in leptin‐treated cells (P < 0.05, unpaired Student's t test) (Fig. 8 A, right). The analysis of the parameters characterizing the shape and size of the quantal events (Carabelli et al. 2007; Gavello et al. 2013) confirmed that there were no significant changes in the kinetic parameters of the spikes (I _max, t _1/2, m, t _p) (Fig. 8 B). Also the catecholamine content (Q) was unaltered and the same was for the cubic root of Q (Q ^1/3), whose mean value is associated with the diameter of chromaffin granules. Q ^1/3 was 0.63 ± 0.02 pC^1/3for the control and 0.64 ± 0.02 pC^1/3 with (leptin 1 nm) (P > 0.1, unpaired Student's t test). Overall, these data demonstrate that leptin (1 nm for 30 min) increases MCC exocytosis without affecting the size of granules and the kinetics of their release.

Figure 8. Amperometric recordings .

A,amperometric spikes were evoked using 30 mm KCl enriched external solution and recorded for 120 s. Representative spikes (indicated by the asterisk) are shown to the right using an expanded time scale. Right: mean estimated values of the mean frequency of the amperometric spikes. This parameter is increased by leptin treatment (*P < 0.05; using Student's unpaired t test). B,main kinetic parameters of the spikes. These are not affected by leptin (1 nm, 30 min).

Discussion

Leptin plays a pivotal role in the control of food intake and energy balance. In addition to its central action, leptin is known to directly affect adrenal CC activity, inducing a two‐ to three‐fold increase of catecholamine release both at rest and during nicotinic stimulation (Takekoshi et al. 1999; Takekoshi et al. 2001 a). In the present study, we have shown that, in cultured MCCs, leptin exerts a dual action on cell excitability and catecholamine release by potentiating the inactivating and non‐inactivating BK currents that shape and regulate the spontaneous and evoked AP firing of these cells (Marcantoni et al. 2010; Martinez‐Espinosa et al. 2014). At rest, leptin dampens the AP firing by silencing the spontaneous activity of the cell, whereas, during sustained stimulation, leptin adapts the cell response with trains of APs of lower frequency but broader duration. During stimulation, leptin also increases the size of the RRP of vesicles and potentiates the frequency of release, thus increasing the quantity of released catecholamines. This is in good agreement with the reported leptin‐mediated increase of catecholamine release in porcine CCs (Takekoshi et al. 1999) and the leptin‐induced increase of circulating catecholamines after stimulation of the hypothalamic‐pituitary‐adrenal axis (Satoh et al. 1999; Harlan et al. 2013; Harlan & Rahmouni, 2013). The effects of leptin are mediated by the ObRs, whose isoforms Ob‐Ra and Ob‐Rb are both expressed in CCs (Hoggard et al. 1997; Takekoshi et al. 1999), as well as by PI3K, which mediates the action of leptin in hippocampal (Shanley et al. 2002 b), hypothalamic (Hill et al. 2008; Yang et al. 2010; Tanida et al. 2015) and parasympathetic neurons (Chae et al. 2005).

Leptin up‐regulates BK channels via PI3K with dual effects on firing at rest and during evoked stimulation

Our data show clearly that leptin acts on MCC excitability by up‐regulating BK channels through the activation of a PI3K‐mediated pathway. At rest, the addition of leptin causes a net hyperpolarization of MCCs with consequent blockade of the spontaneous firing activity. The action is prevented by the BK channel blocker paxilline (1 μm), by the PI3K pathway inhibitor wortmannin and is absent in db^–/db^– mice lacking ObRs (Fig. 1 B–D). Because leptin has no effects on SK channels (Fig. 2 D), the up‐regulation of BK channels appears to be the only cause of cell hyperpolarization and blockage of spontaneous activity during leptin application. The way this occurs if the resting membrane potential in MCCs is mainly regulated by leptin insensitive SK channels (Vandael et al. 2012) (Fig. 2 D) requires clarification. Indeed, BK channels weakly affect the resting membrane potential of MCCs (Marcantoni et al. 2010). Because of their close coupling with Ca_V 1.3 channels, they mainly modulate AP shape and set the firing frequency of MCCs by regulating the after‐hyperpolarization phase of APs (Vandael et al. 2010). However, as shown in Fig. 2 A and B, leptin induces an increase of BK channel conductance and a shift of its voltage‐dependence to more negative potentials, nicely mimicking the effects of BK channel openers (Ha et al. 2006; Bentzen et al. 2014). Thus, BK channels open more readily at resting potentials and can easily drive the cell to more hyperpolarized voltages until firing stops. We have also shown clearly, at macroscopic and single‐channel levels, that the increased open probability of BK channels is not the result of an increased Ca²⁺ concentration from intracellular stores. Indeed, EGTA‐AM does not prevent the actions of leptin (Fig. 3 C) and leptin up‐regulates single BK channels in perforated vesicles kept depolarized to 0 mV to inactivate most Ca_V channels (Fig. 4). In addition, leptin is reported to either preserve the resting intracellular Ca²⁺ in hippocampal neurons (Shanley et al. 2002 a) or cause small transient increases of resting Ca²⁺ concentration in cortical neurons (<15%) (Mancini et al. 2014).

Our present findings are in agreement with previous studies showing that leptin causes hyperpolarization and blockade of AP firing of spontaneously active hippocampal neurons through a PI3K‐driven increased activation of neuronal Sloβ1 BK channels (Shanley et al. 2002 a; Shanley et al. 2002 b), as well as other studies showing that leptin activates BK channels in hippocampus (Gavello et al. 2012) and cortical neurons (Mancini et al. 2014). Shanley et al. 2002 b also reported that leptin and the BK channel opener NS‐1619 equally dampen the induced hyperexcitability of rat hippocampal neurons, in good agreement with the idea that leptin exerts a neuroprotective role during brain damage (Zhang et al. 2007; Signore et al. 2008). It is intriguing, however, that the same is not true for the BK channel openers (Bentzen et al. 2014), probably because the action of leptin as a neuroprotector is more complex and not limited to the potentiation of BK channel activity.

The PI3K‐mediated effect described in the present study is probably a result of leptin‐induced dynamic alterations in the actin cytoskeleton that mediate the activation and synaptic clustering of BK channels (O'Malley et al. 2005). Altered cytoskeleton dynamics may also explain the leptin‐mediated damping of MCC excitability, even if we cannot exclude a direct phosphorylation of the channel by PI3K. Interestingly, in hypothalamic slices, leptin appears to elicit cell depolarization via PI3K activity in POMC neurons. However, leptin withdrawal activates the PI3K pathway in NPY neurons, suggesting that leptin hyperpolarizes NPY neurons via PI3K‐dependent mechanisms (Hill et al. 2008). Indeed, leptin has previously been demonstrated to inhibit hypothalamic cell line excitability by V _m hyperpolarization via K_ATP and PI3K‐dependent mechanisms (Mirshamsi et al. 2004).

An up‐regulation of BK channels also forms the basis of the changes of AP firing during evoked stimulation. Leptin treatment did not alter the cell input resistance and the firing threshold but did lower the initial and steady‐state firing frequency during stronger depolarizations (Fig. 6 A). Leptin‐treated MCCs exhibit faster adaptation to lower firing frequencies, which probably derives from the leptin‐mediated up‐regulation of BK channels. By activating at more negative potentials than the control, the BK channels prolong the after‐hyperpolarization phase, thus allowing the effective recruitment of Na_V and Ca_V channels (Vandael et al. 2015 b). The two channel types are thus able to sustain broader APs of larger amplitude (Fig. 6 B), which is further favoured by the small negative shift of Ca_V channel voltage‐dependence induced by leptin (Fig. 6 D). These changes of firing stabilize the MCC response, ensuring sustained AP trains and increased Ca²⁺‐entry during prolonged stimulation. Thus, in contrast to the effect on resting MCCs, leptin has an adapting action on evoked APs, which, together with the potentiating effects on catecholamine release, furnishes the basis of the sustained action of leptin on sympathetic nerve activity and circulating catecholamines (Satoh et al. 1999) (see below). In relation to this, it is important to emphasize that a study based on electrically evoked AP trains represents a simplification of the depolarizing effects induced by prolonged sympathetic stimulation. Further studies using appropriate patterns of more physiological stimuli, such as those involving nicotinic and muscarinic agonist application, are clearly required. However, previous studies in MCCs have shown that long lasting applications of ACh are dominated by muscarinic receptor activation (Nassar‐Gentina et al. 1988) resulting in sustained membrane depolarizations and increased AP firing similar to those shown in Fig. 6 A.

Leptin increases catecholamine release during stimulation by increasing the RRP and the rate of vesicle release

The effects of leptin on neurosecretion are extremely heterogeneous and depend on the type of endocrine cell or neuron on which the adipokine acts. In pancreatic β‐cells, leptin is reported to either inhibit or potentiate insulin release (Marroqui et al. 2012), whereas, in hypothalamic neurons, leptin inhibits the release of noradrenaline (Brunetti et al. 1999; Kawakami et al. 2008; Kutlu et al. 2010) but stimulates the secretion of luteinizing homrone and follicle‐stimulating hormone (Yu et al. 1997; McCann et al. 1998). This heterogeneous action probably reflects the action of leptin on cell excitability and, thus, on the ion channel types and signalling pathways controlling cell firing, rather than having an effect on the downstream mechanisms regulating trafficking, docking and the fusion of vesicles preceding hormone or neurotransmitter release.

In CCs, there is converging evidence that leptin (1–100 nm) causes an increase of catecholamines synthesis both in bovine and porcine CCs (Takekoshi et al. 2001 b; Shibuya et al. 2002), which is supported by the increase of tyrosine hydroxylase mRNA (Takekoshi et al. 2001 a) via PKC and MAPK‐mediated phosphorylation (Shibuya et al. 2002). An increased synthesis of catecholamines is suggestive of an increased secretion during activity, although clear increments of adrenaline and noradrenaline release have been reported only in porcine CCs (Takekoshi et al. 1999) and PC12 cells (Than et al. 2010). No effects on secretion are reported in bovine (Yanagihara et al. 2000) and human CCs (Glasow et al. 1998), reflecting either the limitations of using different recording techniques or the existence of different cell firing responses to leptin.

In the present study, using capacitance changes and amperometry to detect membrane increases associated to vesicle fusion and catecholamine release, we found that leptin 1 nm (30 min) induces a rise of the overall exocytosis through an augmented size of the RRP and a faster rate of vesicle release in MCCs. The results obtained using both techniques are in good agreement with the findings reported by Takekoshi et al. (1999) and Than et al. (2010). Takekoshi et al. (1999) show that leptin increases the release of adrenaline and noradrenaline regardless of whether porcine CCs are maintained at rest (basal conditions) or depolarized by nicotinic stimulation, suggesting that the potentiating action of leptin on secretion occurs regardless of the mechanisms controlling cell excitability (spontaneous or evoked) and intracellular Ca²⁺ elevation. These findings, similar to ours, indicate an effect of leptin directed toward the mechanism of vesicle secretion. We clearly found that leptin: (1) increases the RRP (Fig. 7); (2) increases the rate of vesicle release (Fig. 8); and (3) preserves the quantity of Ca²⁺ charge, the kinetics of single exocytotic events and the catecholamine content of the granules. This latter finding may appear to be in contrast with the general consensus that leptin increases catecholamine synthesis, although it may be justified considering that an increased catecholamine availability inside the CCs would simply accelerate catecholamines up‐loading on recycled empty vesicles or being used to increase the size of the reserve pool of vesicles that may supply the augmented RRP of vesicles. Further points of convergence of our findings with those of Takekoshi et al. (1999) concern the leptin‐enhanced levels of PKC and cAMP reported in porcine CCs that probably contribute to the increased exocytosis. Indeed, both signalling pathways have been shown to directly potentiate vesicle exocytosis in bovine (Gillis et al. 1996) and rat CCs (Carabelli et al. 2003). Our findings are also in good agreement with those of Than et al. (2010) reporting an enhanced leptin‐mediated secretion of catecholamines in PC12 cells using amperometric detection. Than et al. (2010) report a 50% increase of amperometric spikes during a 2 min KCl‐induced stimulation, which compares well with the 60% increase in the rate of release that we observed in leptin‐treated MCCs obtained under similar stimulation conditions.

Conclusions

Given the evolution in the understanding of leptin physiology, the focus on the role of leptin is now shifted from an anti‐obesity hormone to a weight loss and metabolism regulator. The novelty of our findings is related to the observation that, with a single molecular mechanism involving BK channel activation, leptin induces opposite effects depending on the functional state of the cell. At rest, leptin dampens cell excitability through a deep hyperpolarization of resting membrane potential. By contrast, under stimulation, BK channel activation by leptin has an adapting action on firing activity and elicits an increase in catecholamine secretion. In conclusion, by synthesizing and releasing leptin, the adipose tissue is able to modulate the electrical activity and catecholamine release of CCs, thus playing a pivotal role in metabolism regulation.

Additional information

Competing interests

The authors declare that they have no competing interests.

Author contributions

DG contributed to the data collection of the current clamp and capacitance recordings, as well as part of the voltage clamp recordings. DV contributed to voltage clamp, current clamp and single‐channel experiments. SG performed the amperometric experiments. DG, DV, EC and VC contributed to the conception and design of the experiments and the drafting of the manuscript. EC and VC revised the manuscript critically, giving important suggestions for improvement. All authors have approved the final version of the manuscript submitted for publication.

Fundings

This work was supported by the Compagnia di San Paolo Foundation ‘Neuroscience Program’ to VC and ‘Progetto di Ateneo 2011‐13’ to EC.